CN105379212B

CN105379212B - Method, controller and system for managing electronic content ingestion in a content distribution network

Info

Publication number: CN105379212B
Application number: CN201480041255.0A
Authority: CN
Inventors: C.菲利普斯; J.罗杰斯; J.雷诺兹; B.富尔斯曼; C.达舍
Original assignee: Ericsson Television Inc
Current assignee: Ericsson Television Inc
Priority date: 2013-05-20
Filing date: 2014-05-19
Publication date: 2021-05-07
Anticipated expiration: 2034-05-19
Also published as: EP3000210A4; WO2014188328A1; US20140344879A1; CN105379212A; US8973073B2; EP3000210A1

Abstract

An ingest controller, system and method for managing ingestion of electronic content in a content distribution network (CDN) (11). Content is received from one or more content sources (CS) (12‑15) through the inbound content pipeline (19) with maximum bandwidth capacity. The database stores the bandwidth weights assigned to each CS by the Network Policy Management System (PMS) (46). The bandwidth weight assigned to each CS corresponds to a fraction of the maximum bandwidth capacity of the inbound content pipeline. The input buffer (41) receives content from the CS in one or more input buffer queues (51-54) associated with the CS. The content fetching mechanism fetches content from the input queue at each CS's upload rate based at least on each CS's assigned bandwidth weight, thereby enabling a flow control mechanism (55) to control the upload of each CS in proportion to the bandwidth weight bit rate.

Description

Method, controller and system for managing electronic content ingestion in a content distribution network

Technical Field

The present disclosure relates to methods and apparatus for weighted policy management for ingestion of video or audio content in a Content Delivery Network (CDN).

Background

Network bandwidth is often crowded with multiple download or upload sessions competing for network resources. In existing file delivery systems (e.g., CDNs), files are uploaded from a content source to a "watch folder" for further delivery over an operator network (which is not owned or controlled by the content source or CDN owner). In such systems, content upload from opposing content sources is competing, often resulting in sub-optimal bandwidth distribution (e.g., one content source may utilize a disproportionate share of the bandwidth, thereby starving the other content sources). Due to the lack of control, CDN owners cannot guarantee quality levels to content sources on different service layers.

Fig. 1 is a simplified block diagram of an existing network architecture 10 for uploading content to a CDN 11. The content sources CS-1 through CS-412-15 may compete to upload content over the Internet 16 to a watch folder disk farm 17(watch folder disk farm) in a data center 18 commonly accessible by the content sources and CDN. In the illustrated example, CS-1 to CS-3 each upload associated content using an Internet connection having a maximum bandwidth of 100 Mbs. CS-4 utilizes an internet connection with a maximum bandwidth of 30 Mbs. The inbound content virtual pipe 19 from the internet to the monitoring file chuckhouse has a maximum bandwidth capacity of 200Mbs in the example shown. The inbound content is first received in the data center 18 in the input buffer 20, then retrieved and written to the monitoring file chuck field 17. From monitoring the file chucky, a CDN delivery system 21 delivers ingested content to a CDN 11, which may include an origin server 22 and a plurality of regional servers 23.

Each of the content sources 12-15 is programmed to: when the content source has content to upload, grab up to the currently available bandwidth, up to the maximum bandwidth of its internet connection. Thus, the maximum bandwidth capacity of the inbound content pipe 19 is exceeded when more than two of the content sources desire to upload content simultaneously in this example. Any other content source that attempts to upload content to the watch folder will then be blocked. Thus, the system is essentially first come first serve. For example, if CS-1 and CS-2 are uploading content, CS-3 and CS-4 must wait until CS-1 or CS-2 completes its upload or pauses in its upload for a sufficient time for CS-3 or CS-4 to jump in and take its place. Of course, this can cause additional problems if a content source partially uploads a content file and then cannot re-access the watch file cartridge field 17 because network bandwidth is seized by another content source.

Disclosure of Invention

The present disclosure provides a system that manages the uploading of third party content to a watch folder of a CDN for further distribution. Based on the weighted pre-established policy, an ingestion controller in the CDN controls the speed for uploading content from various content sources. When multiple content sources contend for network bandwidth and the total required bandwidth is greater than the maximum bandwidth capacity of the inbound content pipe, the ingestion controller may limit the upload speed of each content source such that the total ingestion bandwidth utilized by the content sources is equal to the maximum bandwidth capacity of the inbound content pipe. The ingestion controller may allow the requesting content source to upload its content at normal speed when the demand for network bandwidth drops to a level equal to or less than the maximum bandwidth capacity of the inbound content pipe.

It is an object of the present disclosure to minimize or avoid any modifications to the content source. Thus, in one exemplary embodiment, the ingestion controller may limit the upload speed of the content sources by, for example, limiting the speed at which content is retrieved from the associated input buffer queue associated with each content source. The acquisition rate from each buffer queue is based at least on the assigned bandwidth weighting of each associated content source. In one embodiment, the ingestion controller limits the disk write speed that allows the Operating System (OS) monitoring the folders to fetch the contents of each source from its associated input buffer queue and write the contents to memory in the monitoring folder. This causes the input buffer queues associated with the content sources to become full, further causing flow control mechanisms in the transmission network, e.g., the internet, to control the flow of packets from each content source. When the demand for network bandwidth drops to a level equal to or less than the maximum bandwidth capacity of the inbound content pipe, the ingestion controller may allow the OS to write to memory at normal speed, thereby allowing the content source to send its content at its full rate. During network contention time, advanced users may be allocated a faster disc write speed than non-advanced users. All users on a particular service layer may be assigned equal disc writing speeds.

In one embodiment, a method for managing ingestion of electronic content in a CDN is disclosed. Content is received from one or more content sources through an inbound content pipe having a maximum bandwidth capacity. The method comprises the following steps: obtaining, by the ingestion controller, a bandwidth weighting assigned to each of the one or more content sources, wherein the bandwidth weighting assigned to each content source corresponds to a fraction of a maximum bandwidth capacity of the inbound content pipe; and controlling, by the ingestion controller, an upload rate from each content source based at least on the assigned bandwidth weighting for each content source.

The ingestion controller may include an input buffer connected to the memory, wherein the input buffer is configured to receive content from one or more content sources in one or more input buffer queues associated with the content sources, and the memory is configured to store the received content upon retrieval from the input buffer. The controller may control the upload rate by controlling the speed at which content is retrieved from each queue of the input buffer based at least on the assigned bandwidth weighting for each associated content source. Slowing the acquisition speed of a given input queue causes the given input queue to become full, thereby triggering a network flow control mechanism that causes the content source associated with the given input queue to slow its transmission rate as instructed from the flow control mechanism.

In another embodiment, an ingestion controller for managing ingestion of electronic content in a CDN is disclosed. Content is received from one or more content sources through an inbound content pipe having a maximum bandwidth capacity. The ingestion controller includes a database of bandwidth weightings assigned to each of the one or more content sources, wherein the bandwidth weighting assigned to each content source corresponds to a fraction of a maximum bandwidth capacity of the inbound content pipe. The controller further comprises: an input buffer configured to receive content from one or more content sources in one or more input buffer queues associated with the content sources; and a content retrieval mechanism configured to receive the bandwidth weightings from the database and retrieve content from the input queue at an upload rate for each content source based at least on the assigned bandwidth weighting for each content source.

In another embodiment, a system for managing ingestion of electronic content in a CDN is disclosed. Content is received from one or more content sources through an inbound content pipe having a maximum bandwidth capacity. The system includes an operator Policy Management System (PMS) configured to assign bandwidth weightings to one or more content sources, wherein the bandwidth weighting assigned to each content source corresponds to a fraction of a maximum bandwidth capacity of the inbound content pipe. The system also includes an ingestion controller configured to control an upload bit rate of each of the one or more content sources. The ingestion controller includes: a database configured to store bandwidth weightings assigned to each of one or more content sources; an input buffer configured to receive content from one or more content sources in one or more input queues associated with the content sources; and a content retrieval mechanism configured to receive the bandwidth weightings from the database and retrieve content from the input queue at an upload rate for each content source based at least on the assigned bandwidth weighting for each content source.

The system enables CDN owners to provide high-level services to content sources by providing a prioritized upload speed to them. In addition, the system prevents content sources from utilizing a disproportionate share of bandwidth at the expense of other content sources at the same layer. Other features and benefits of embodiments of the present invention will become apparent from the detailed description below.

Drawings

In the following part, the invention will be described with reference to exemplary embodiments shown in the drawings, in which:

FIG. 1 (Prior Art) is a simplified block diagram of an existing network architecture for uploading content to a CDN;

FIG. 2 is a simplified block diagram of an exemplary embodiment of a network architecture including an ingestion controller for controlling the uploading of content to a CDN;

FIG. 3 is a flow diagram of an exemplary embodiment of an overall method for controlling the upload of content to a CDN;

FIG. 4 is a simplified block diagram of an exemplary embodiment of the ingestion controller of FIG. 2;

FIG. 5 is a simplified block diagram of an exemplary embodiment of input buffer 41 of FIG. 4;

FIG. 6 is an exemplary embodiment of a Weighted Fair Queuing (WFQ) weight table;

FIG. 7 is a flow diagram illustrating in greater detail an exemplary embodiment of a method for controlling the upload of content from two or more content sources to a CDN;

FIG. 8 is a block diagram of an exemplary embodiment of a nested data structure in an approval framework for writing received content to a watch folder, thereby limiting the rate of retrieval from an input buffer; and

FIG. 9 is a flow diagram that schematically illustrates an embodiment of a method for writing a thread (lightweight process) to a monitor folder.

Detailed Description

The present invention now will be described more fully hereinafter with reference to the accompanying drawings, in which preferred embodiments of the invention are shown. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. In the drawings, like reference numerals designate like elements. Additionally, it should be understood that the present invention can be implemented in hardware or a combination of hardware and software stored on a non-transitory memory and executed by a general purpose computer or microprocessor.

Fig. 2 is a simplified block diagram of a network architecture 25 for uploading content to CDN 11, in accordance with certain embodiments. Content sources CS-1 through CS-412-15 may compete to upload content over a packet data network, such as the internet 16, to a surveillance file chucker farm 17 in a data center 18 commonly accessible by the content sources and CDNs. It should be understood that in the context of the present disclosure, the term "disk farm" is not intended to limit monitoring folder functionality to a particular hardware implementation. The monitoring folder may be any suitable memory device with sufficient memory capacity and response time to store content from multiple content sources before forwarding the content to the CDN. For example, the memory device may be one or more of the following: CPU registers, cache on CPU, Random Access Memory (RAM), cache on disk, Compact Disk (CD), Solid State Disk (SSD), in-network (cloud) storage, tapes, and other backup media. In the remainder of the description herein, the watch folder functionality will be referred to simply as a "watch folder".

In the illustrated example, CS-1 to CS-3 each upload associated content using an Internet connection having a maximum bandwidth of 100 Mbs. CS-4 utilizes an internet connection with a maximum bandwidth of 30 Mbs. The inbound content virtual pipe 19 from the internet to the monitoring folder 17 has a maximum bandwidth capacity of 200Mbs in the illustrated example. From the monitoring folder, CDN delivery system 21 delivers ingested content to CDN 11, which may include origin server 22 and a plurality of regional servers 23.

When multiple content sources contend for network bandwidth and the total required bandwidth is greater than the maximum bandwidth capacity of the inbound content pipe 19, the ingestion controller 26 is configured to control the upload speed of each content source such that the total ingestion bandwidth utilized by the content source is equal to or less than the maximum bandwidth capacity of the inbound content pipe. The ingestion controller may allow the requesting content source to upload its content at normal speed when the demand for network bandwidth drops to a level equal to or less than the maximum bandwidth capacity of the inbound content pipe. The operation of the ingestion controller is described in more detail below in conjunction with fig. 4.

Fig. 3 is a flow diagram of an exemplary embodiment of an overall method for controlling the upload of content from two or more content sources 12-15 to a CDN 11. At step 31, the ingestion controller 26 obtains a bandwidth weighting 42 assigned to each of the two or more content sources, where the bandwidth weighting assigned to each content source corresponds to a fraction of the maximum bandwidth capacity of the inbound content pipe 19. In step 32, the ingestion controller 26 controls the upload bit rate from each content source based at least on the assigned bandwidth weighting 42 for each content source.

Fig. 4 is a simplified block diagram of an exemplary embodiment of the ingestion controller 26 of fig. 2. In the illustrated embodiment, the ingestion controller 26 includes: an input buffer 41 comprising a flow control mechanism 55; a set of predefined Weighted Fair Queuing (WFQ) weights 42 associated with the content sources 12-15; and a disk writer 43 that monitors an Operating System (OS) of the folder 17. The operation of the ingestion controller may be controlled, for example, by a processor 44 running computer program instructions stored on a memory 45.

The WFQ weights 42 include bandwidth weightings assigned to each of the content sources 12-15 by the operator Policy Management System (PMS)46 based on network policies and the service layer of the respective content source. Each WFQ weight corresponds to a fraction of the maximum bandwidth capacity of the inbound content pipe 19 allocated to the associated content source. The PMS may dynamically assign WFQ weights to content sources as the contention level for network bandwidth increases and decreases. As described in more detail below, each content source may also assign weights to individual content streams within the content source's allocated bandwidth.

Fig. 5 is a simplified block diagram of an exemplary embodiment of the input buffer 41 of fig. 4. Input buffer 41 may be configured to receive content from content sources in a plurality of input queues 51-54, where each input queue is associated with a different one of content sources CS1-CS 412-15. The size of each queue may be determined in proportion to the allocated bandwidth of the associated content source. The flow control mechanism 55 interacts with the content sources to control the upload bit rate of each source using flow control instructions 56, as described in more detail below.

Fig. 6 is a functional block diagram of an exemplary embodiment of a table of WFQ weights 42 applied to an input buffer 41 for controlling bandwidth utilization of various content sources. The left column shows WFQ weights that may be applied by the PMS 46 to each content source according to network policy and the service layer of the respective content source. Without weight, each content source would attempt to transmit at its maximum rate, requiring an inbound content pipe with a bandwidth of 330 Mbs. However, the maximum capacity of the inbound content pipe is 200Mbs, so the weight must reduce the maximum transmission rate to the allowable transmission rate, the sum of which does not exceed 200 Mbs. In the illustrated example, CS-1 and CS-2 are assigned a rank "3 weight" (66.667Mbs), CS-3 is assigned a rank "2 weight" (44.444Mbs), and CS-4 is assigned a rank "1 weight" (22.222 Mbs). If a content source stops transmitting, or if another content source starts transmitting, the WFQ weights may be adjusted up or down to keep the sum of the transmission rates within 200Mbs capacity.

The content source may upload multiple content streams within the content source's allocated bandwidth. Such a multi-stream content source may also allocate bandwidth to individual content streams according to priorities assigned by the multi-stream content source. For example, CS-112 may have three content streams 61-63 to upload to the CDN, and may divide its 66.667Mbs among the three streams. As shown in this example, CS-1 assigns

priority levels

1, 2, and 3 to the three content streams that are associated with bandwidths of 33.334Mbs, 22.222Mbs, and 11.111Mbs, respectively. The multi-stream content sources CS-2 and CS-3 also assign varying priority levels to their multiple content streams.

Fig. 7 is a flow diagram illustrating in more detail an exemplary embodiment of a method for controlling the upload of content from two or more content sources 12-15 to a CDN 11. At step 71, the PMS 46 assigns a bandwidth weight to the CS with the content to be uploaded to the CDN 11, for example. The weights may be preconfigured and assigned to the network connections of the CS and the content flowing through these connections before the content is actually transferred. The weights may also change during transmission. At step 72, the CS starts uploading its content. At step 73, the OS/disk writer 43 retrieves the content from the input queues 51-54 of the input buffer 41 according to the assigned bandwidth weights and writes the content to the watch folder 17. The ingest controller causes the OS/disk writer to reduce its speed of fetching content from the buffer queue of each content source in proportion to the WFQ weights 42 assigned to each content source and each individual stream (if applicable). The bandwidth weight does not affect the upload bit rate unless there is a competing content transmission that exceeds or is expected to exceed the maximum bandwidth capacity of the inbound content virtual pipe 19. Only then does the intake controller 26 start controlling the intake speed.

Even if each content source desires to transfer content at the maximum bit rate of its connection to the internet 16, the present disclosure controls the average upload bit rate of each content source with the speed at which the OS/disk writer 43 retrieves content from the buffer queues 51-54 and writes the content to the monitoring folder 17 so that the maximum bandwidth capacity of the inbound content pipe 19 is not exceeded. When the OS/disc writer 43 slows down the acquisition speed of a particular content source, the queue in the input buffer 41 associated with that particular content source fills up immediately. At step 74, the flow control mechanism 55 associated with the input buffer is triggered to slow down the upload bit rate of the associated content source. The flow control mechanism may be a standard flow control mechanism such as the Transaction Control Protocol (TCP) used for signaling over the internet and other Internet Protocol (IP) based networks. The flow control instructions 56 may be included in acknowledgement messages sent from the input buffer 41 to the various content sources. The flow control instructions may indicate, for example, that the content source in question is authorized to transmit a given number of data packets, but then must wait for further authorization before transmitting additional packets. When the content source completes its upload or a new content source is added, step 75, the method returns to step 71 where the sum is recalculated and then re-evaluated, step 72.

In this way, by controlling the buffer acquisition rate of each of the content sources 12-15, the ingestion controller 26 controls the amount of network bandwidth utilized by each content source. When the buffer retrieval speed of a particular content source is further limited, the network bandwidth utilization of that content source is reduced because the input queue of that source remains full for a longer period of time, and the flow control mechanism 55 also limits the number of packets permitted to be transmitted by that particular content source. During network contention time, advanced users may be allocated faster buffer acquisition and disk write speeds than non-advanced users. All users on a particular service layer may be assigned equal buffer acquisition and disc write speeds.

FIG. 8 is a block diagram of an exemplary embodiment of a nested data structure in an approval framework for writing received content to a watch folder, thereby limiting the rate of retrieval from an input buffer.

FIG. 9 is a flow diagram that schematically illustrates an embodiment of a method of writing a thread (lightweight process) to a monitoring folder 17 (e.g., RAM, disk, etc.).

Fig. 8 and 9 collectively illustrate an exemplary method of nesting data structures used to determine which data flows to service and the times at which they are to be serviced. Referring to FIG. 8, a file writer thread 81 loads a token 82 into a leaf sequencer (leaf sequencer) 83. The leaf sequencer holds the token from the write thread. In various embodiments, the leaf sequencer may represent a separate device in the subscriber's home or a video on demand asset from a content source. A Weighted Fair Queuing (WFQ) sequencer 84 pulls tokens from the leaf sequencers and applies a weighted write queuing algorithm to the weighted file write queues 51-54. In various embodiments, a WFQ sequencer may be used to divide bandwidth among various devices within a subscriber's home, or among video on demand assets from a content source. The weighted write queuing algorithm is described in more detail below.

A set of rate limiters 85 control the input rate into the weighted file write pipe queue 86. In various embodiments, rate limiter 85 may represent a home or content provider. The common WFQ sequencer 86, shown as a weighted file write pipe queue, also applies a weighted write queuing algorithm to the data. The WFQ sequencer 86 has internal logic that selects one of the rate limiters 85 to pull a token from. A final rate limiter 87, having a different value than rate limiter 85, takes tokens from WFQ sequencer 86 and delays the tokens to keep the data rate below its configured limit. The grant loop pulls tokens from the rate limiter 87 and grants them. The rate limiter 87 essentially calculates the time when the content packet should be transmitted by the CS. When there is no congestion in the inbound content virtual pipe 19, the rate limiter calculates the past transmission time, thus approving the immediate transmission. When there is network congestion in the virtual pipe, the rate limiter calculates a future transmission time, thus delaying transmission approval.

It should be noted that the rate limiter, WFQ sequencer, and leaf sequencer can be set to a number of different configurations to model different business needs.

Referring to FIG. 9, a single thread write process is shown, with multiple threads being handled by the loop shown. At step 91, the disk writer 43 monitoring the folder OS generates a list of bytes of a single thread to be written to memory. At step 92, the OS generates an N byte token 82, where N is equal to the byte block size. At step 93, the token for the thread is loaded into one of the approval frameworks, i.e., leaf sequencers 83. At step 94, the thread waits for its token to be granted. Once the token is approved, N bytes are written to memory at step 95. At step 96, the OS determines whether there is still a payload to be transferred. If not, the method moves to step 97 where the file descriptor is closed. If there is still payload present, the method loops back to step 92 and the process repeats for another thread.

In one embodiment, the weighted write queuing algorithm may be expressed by:

x_i＝(A_i+Tb_i)/w_i

wherein:

x_iis an intermediate value used to account for (account) queue i in the hypothetical bitwise cyclic model, which represents the estimated completion time of the transmission;

A_i＝sum(Qa_ik) Wherein Qa_ikIs from the write queue Q_iThe amount of bytes of the previously approved token k;

Tb_ifrom write queue Q_iThe number of bytes of the current token; and

w_iwrite queue Q_iThe weighting factor of (2).

Thus, A_iIs the sum of the bytes of queue i that have been transmitted within the "specification window". A. the_i+Tb_iAnd adding the byte number of the candidate token. Weight w_iCorresponding to the "speed" of the transmit queue relative to other weighted streams.

By comparing x of various queues_iThe value can determine which packet will complete the transmission first in the hypothetical bitwise round robin scheme. That packet is selected for transmission. Approval from queue Q_iA token of (2), wherein x_i＝min(x_i). The (reconcile) tokens are reconciled by weight.

In the drawings and specification, there have been disclosed typical preferred embodiments of the invention and, although specific terms are employed, they are used in a generic and descriptive sense only and not for purposes of limitation.

Claims

1. A method for managing ingestion of electronic content in a content delivery network, CDN, (11), wherein the content is received from one or more content sources (12-15) through an inbound content pipe (19) having a maximum bandwidth capacity, the method comprising:

obtaining, by an ingestion controller (26), a bandwidth weighting (42) assigned to each of the one or more content sources, wherein the bandwidth weighting assigned to each content source corresponds to a fraction of a maximum bandwidth capacity of the inbound content pipe (19); and

controlling, by the ingestion controller (26), upload rates from the content sources based at least on the assigned bandwidth weights (42) of the content sources such that a sum of the upload rates from the content sources is maintained within a maximum bandwidth capacity of the inbound content pipe (19); wherein the ingestion controller (26) comprises an input buffer (41) connected to a memory (17), wherein the input buffer is configured to receive the content from the one or more content sources (12-15) in one or more input buffer queues (51-54) associated with the content sources, and the memory is configured to store the received content for distribution to a CDN (11) upon retrieval from the input buffer, wherein each input buffer queue is associated with a different one of the one or more content sources, and wherein the controlling comprises:

controlling a speed of fetching content from respective input buffer queues (51-54) of the input buffer (41) based at least on an assigned bandwidth weighting (42) of respective associated content sources, wherein slowing down a fetching speed of a given input buffer queue causes the given input buffer queue to become full, thereby triggering a network flow control mechanism (55), the network flow control mechanism (55) causing a content source associated with the given input buffer queue to slow down its transmission rate as instructed from the network flow control mechanism (55).

2. The method of claim 1, wherein when two or more content sources are on the same service layer, the controlling comprises:

assigning, by an operator policy management system PMS (46), a same bandwidth weighting (42) to each of the two or more content sources (12-15); and

allocating, by the ingestion controller (26), an equal number of shares of a maximum bandwidth capacity of the inbound content pipe to each of the two or more content sources.

3. The method of claim 2, wherein when the one or more content sources are on different service layers, the controlling comprises:

assigning, by the operator PMS (46), a different higher bandwidth weighting (42) to a given content source at a higher service level than the rest of the one or more content sources; and

allocating, by the ingestion controller (26), a greater fraction of the maximum bandwidth capacity of the inbound content pipe to the given content source on the higher service layer.

4. The method of claim 3, further comprising obtaining, by the ingestion controller (26), category weightings (61-63) for individual content streams from a multi-stream content source (12), wherein the controlling comprises controlling upload rates of individual content streams from the multi-stream content source based on: the assigned bandwidth weightings of the multi-stream content sources and the assigned category weightings of the individual content streams.

5. The method of claim 4, wherein the category weightings (61-63) are assigned by the multi-stream content source (12) based on one or more of:

a topic of each content stream;

the release date of each content stream;

a distribution frequency of the portions of each content stream; and

the projected demand for each content stream.

6. An ingestion controller (26) for managing ingestion of electronic content in a content delivery network, CDN, (11), wherein the content is received from one or more content sources (12-15) through an inbound content pipe (19) having a maximum bandwidth capacity, the ingestion controller (26) comprising:

a database of bandwidth weightings (42) assigned to each of the one or more content sources (12-15), wherein the bandwidth weighting assigned to each content source corresponds to a fraction of a maximum bandwidth capacity of the inbound content pipe (19);

an input buffer (41) configured to receive the content from the one or more content sources (12-15) in one or more input buffer queues (51-54) associated with the content sources, wherein each input buffer queue is associated with a different one of the one or more content sources; and

a disk writer (43) configured to receive the bandwidth weightings (42) from the database and to retrieve content from the input buffer queues (51-54) at an upload rate for each content source based at least on the assigned bandwidth weighting (42) for each content source such that a sum of the upload rates from each content source is maintained within a maximum bandwidth capacity of the inbound content pipe (19) and to store the received content in memory for distribution to the CDN (11) upon retrieval from the input buffer queues (51-54), and wherein the controlling comprises:

7. The ingestion controller (26) as recited in claim 6, wherein the database is configured to receive the bandwidth weighting from an operator policy management system, PMS, (46), wherein when two or more content sources are on a same service layer, the PMS assigns a same bandwidth weighting (42) to each of the two or more content sources; and

wherein, when assigning the same bandwidth weighting (42) to each of the two or more content sources, the disc writer (43) is configured to retrieve content from the input buffer queues (51-54) of each of the two or more content sources at the same upload rate.

8. The ingestion controller (26) as recited in claim 7, wherein when there are two or more content sources and a given content source is on a higher service layer than the rest of the content sources, the PMS (46) assigns a different higher bandwidth weighting (42) to the given content source on the higher service layer; and

wherein the disc writer (43) is configured to retrieve content from an input buffer queue of the given content source at a higher upload rate.

9. The ingestion controller (26) as claimed in claim 8, wherein the database is further configured to receive a category weighting (61-63) for each individual content stream from a multi-stream content source (12) transmitting more than one content stream, wherein the disc writer (43) is configured to fetch content from the input buffer queue for each individual stream based on: the assigned bandwidth weightings of the multi-stream content sources and the assigned category weightings of the individual content streams.

10. The ingestion controller (26) as claimed in claim 9, wherein the category weightings (61-63) are assigned by the multi-stream content source (12) based on one or more of:

a topic of each content stream;

the release date of each content stream;

a distribution frequency of the portions of each content stream; and

the projected demand for each content stream.

11. A system (25) for managing ingestion of electronic content in a content delivery network, CDN, (11), wherein the content is received from one or more content sources (12-15) through an inbound content pipe (19) having a maximum bandwidth capacity, the system comprising:

an operator policy management system, PMS, (46) configured to assign a bandwidth weighting (42) to the one or more content sources (12-15), wherein the bandwidth weighting assigned to each content source corresponds to a fraction of a maximum bandwidth capacity of the inbound content pipe (19); and

an ingestion controller (26) configured to control an upload rate of each of the one or more content sources, the ingestion controller comprising:

a database configured to store the bandwidth weightings (42) assigned to each of the one or more content sources (12-15);

a disk writer (43) configured to receive the bandwidth weighting (42) from the database, and retrieving content from the input buffer queues (51-54) at an upload rate for each content source based at least on the assigned bandwidth weighting (42) for each content source, such that the sum of the upload rates from the content sources is maintained within the maximum bandwidth capacity of the inbound content pipe (19), wherein the input buffer (41) comprises a network flow control mechanism (55), wherein a given input buffer queue becomes full when the disk writer (43) slows down the fetching speed of the given input buffer queue, and triggering the network flow control mechanism (55) to send an instruction to a content source associated with the given input buffer queue to slow its transmission rate in accordance with the instruction from the network flow control mechanism (55).

12. The system (25) of claim 11, wherein the database is configured to receive the bandwidth weighting from an operator policy management system, PMS, (46), wherein when two or more content sources are on the same service layer, the PMS assigns the same bandwidth weighting (42) to each of the two or more content sources; and

wherein the disk writer (43), when assigning the same bandwidth weighting (42) to each of the two or more content sources, is configured to retrieve content from the input buffer queues (51-54) of each of the two or more content sources at the same upload rate.

13. The system (25) of claim 12, wherein when there are two or more content sources and a given content source is on a higher service layer than the rest of the two or more content sources, the PMS (46) assigns a different higher bandwidth weighting (42) to the given content source on the higher service layer; and

14. The system (25) of claim 13, wherein the database is further configured to receive a category weighting (61-63) for each individual content stream from a multi-stream content source (12) that transmits more than one content stream, wherein the disk writer (43) is configured to retrieve content from the input buffer queue for each individual stream based on: the assigned bandwidth weightings of the multi-stream content sources and the assigned category weightings of the individual content streams.

15. The system (25) of claim 14, wherein the category weightings (61-63) are assigned by the multi-stream content source (12) based on one or more of:

a topic of each content stream;

the release date of each content stream;

a distribution frequency of the portions of each content stream; and

the projected demand for each content stream.